Abstract
Real-world statistical regularities, or expectations about the visual world learned over a lifetime, have been found to be associated with perceptual efficiency. One example of a real-world statistical regularity is that good (i.e., highly representative) exemplars of basic scene categories, such as cities and mountains, are detected more readily than bad exemplars of the category (Caddigan et al., 2017). Similarly, good exemplars are more accurately decoded than bad exemplars in scene-responsive regions, especially the parahippocampal place area (PPA) (Torralbo et al., 2013). Here we used an event-related fMRI design to ask whether the observed neural advantage of the good scene exemplars requires full attention. Attention was directed away from the scenes with a distracting Rapid Serial Visual Presentation (RSVP) task superimposed on the center of the scenes. We used good and bad exemplars of mountains and cities from Torralbo et al. (2013). To assess the clarity of the neural representation of category we asked how well scene category was decoded with a Support Vector Machine (SVM) and how category cohesion (within-category representational similarity) and category distinctiveness (between-category representational dissimilarity) differed for good and bad exemplars. In the attend-to-scenes condition, our results replicated an earlier study showing that good exemplars not only evoked less activity, suggesting more efficient processing, but also a clearer category representation (better decoding and higher cohesion and distinctiveness) than bad exemplars. Importantly, similar advantages of the good exemplars were also observed when participants were distracted by the demanding RSVP task. In addition, a cross-decoding method between attended and distracted representations revealed that attention resulted in a quantitative rather than qualitative improvement of the category representation, particularly for good exemplars. We therefore conclude that the advantage of good exemplars on neural representations does not require full attention.